A method to determine rates and patterns of variability in ecological communities

It is well known that ecological communities are spatially and temporally d ynamic. Quantifying temporal variability in ecological communities is chall enging, however, especially for time-series data sets of less than 40 measu rement intervals. In this paper, we describe a method to quantify temporal variability in multispecies communities over time frames of 10-40 measureme nt intervals. Our approach is a community-level extension of autocorrelatio n analysis, but we use Euclidean distance to measure similarity of communit y samples at increasing time lags rather than the correlation coefficient. Regressing Euclidean distances versus increasing time lags yields a measure of the rate and nature of community change over time. We demonstrate the m ethod with empirical data sets from shortgrass steppe, old-field succession and zooplankton dynamics in lakes, and we investigate properties of the an alysis using simulation models. Results indicate that time-lag analysis pro vides a useful quantitative measurement of the rate and pattern of temporal dynamics in communities over time frames that are too short for more tradi tional autocorrelation approaches.